Re: bugzilla (Was: What do you want me to do?) - Mailing list pgsql-hackers

From Andrew Dunstan
Subject Re: bugzilla (Was: What do you want me to do?)
Date
Msg-id 3FB0F529.7010005@dunslane.net
Whole thread Raw
In response to Re: bugzilla (Was: What do you want me to do?)  (Peter Eisentraut <peter_e@gmx.net>)
Responses Re: bugzilla (Was: What do you want me to do?)
List pgsql-hackers
Peter Eisentraut wrote:

>Andrew Dunstan writes:
>
>  
>
>>Seriously, I have wondered if it might be a good idea to assemble a
>>small "hit team" that would take some high profile open source projects
>>and make sure they worked with Postgres. Bugzilla would be the most
>>obvious candidate, but there are certainly others. I suspect that could
>>be quite productive, though.
>>    
>>
>
>Good thought, but a hit team is not the right answer, because any project
>that would have been "hit" in this way will just go bad again the moment
>its database layer is changed.  What would work better are "consultants":
>people that hang around on the other project's mailing lists, offer advise
>on database layer modelling and implementation, do clean up tasks, check
>regularly if everything works with the PG development branch, be there
>when the developers of that other project have a question.  I've been
>doing a bit of that, and my sensation is that most developers of
>database-backed applications are dying to have people like that at their
>disposal.
>
>  
>
Well, I didn't mean "hit and run" team ;-). I take your point, though. 
However, some projects will need a big effort up front - Bugzilla 
certainly will. I am doing some work on it but any help will be 
appreciated. Part of the problem is that there is no layering of the 
app, separating out the logical and physical views. So instead of a nice 
procedural layer where one could concentrate the creation of Db-specific 
SQL, it is littered throughout the app. Some of the issues include:
. efficient replacement of the enumerated types
. transactional safety
. properly replacing the calles to "replace into"
. full text searching
. migration scripts
As you can see, doing it properly is quite a big job.

However, I also think that there is value in identifying those projects 
that will give the best bang for the buck for our project, and then 
trying to concentrate some resources on those. Your suggestion elsewhere 
of "pick your second favourite app" is likely to result in a more 
scattergun approach. Also, if it had the imprimatur of the PostgreSQL 
community to some extent appraoches to projects might be more welcome - 
"Dear open-source-project-manager, on behalf of the PostgrSQL community 
we would like to offer you assistance in making sure your application 
works with PostgrSQL, the world's most advanced open-source database 
system...."

cheers

andrew



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: bug in 7.4 ...
Next
From: Shridhar Daithankar
Date:
Subject: Re: bugzilla (Was: What do you want me to do?)